Financial services

Chromatography

Support vector machine

Accounting

Computer science

Business

Insurance fraud

Naive Bayes classifier

Financial statement

Audit

Data science

Artificial intelligence

Chemistry

Sample (material)

Finance

Financial fraud

Computer Science Review

This paper gives a comprehensive revision of the state-of-the-art research in detecting financial fraud from 2009 to 2019 inclusive and classifying them based on their types of fraud and data mining technology utilized in detecting financial fraud. The review result yielded a sample of 75 relevant articles (58 conference papers with 17 peer-reviewed journal articles) that are categorized into four main groups (bank fraud, insurance fraud, financial statement fraud, and cryptocurrency fraud). The study shows that 34 data mining techniques were used to identify fraud throughout various financial applications. The SVM is found to be one of the most widely used financial fraud detection techniques that carry about 23% of the overall study, followed by both Naïve Bayes and Random Forest, resulting in 15%. The results of our comprehensive review revealed that most data mining techniques are extensively implemented to bank fraud and insurance fraud with a total of 61 research studies out of 75 that constitute the largest portion equal to 81.33% of the overall number of papers. This review provides a good reference source in guiding the detection of financial fraud for both academic and practical industries with useful information on the most significant data mining techniques used and shows the list of countries that are exposed to financial fraud. Our review contributes by expanding the sample of the reviewed articles that were not included by previous research and presents a summary of the prominent works done by various researchers in the field of financial fraud.

Financial fraud detection applying data mining techniques: A comprehensive review from 2009 to 2019